drafthorse 2025.2.0

Python ZUGFeRD XML implementation

This is a low-level python implementation of the ZUGFeRD XML format. ZUGFeRD is a German format for sending digital invoices. ZUGFeRD XML files are to be attached to a PDF file. This library can be used to generate or parse the contents of this XML file as well as attach it to a PDF. We do not support parsing PDF files (for now).

By low-level, we mean that this library tries to model the ZUGFeRD data model 1:1 without any further abstractions or simplifications. You can set and parse all parameters defined in ZUGFeRD 2.3. To see which parameters are defined, please refer to the official ZUGFeRD specification, there is no documentation for this library that reiterates it.

All output is validated against the official XSDs, but no validation of profile levels (basic, comfort, extended) is performed.

The profile level is detected automatically based on the XML data and added to the PDF metadata.

Maintenance note

With the possibly upcoming legal obligation to use digital formats for all B2B invoices in Germany, interest in this library is increasing. We built this library to ble able to use it in the [pretix-zugferd](https://github.com/pretix/pretix-zugferd) module and we are planning to keep maintaining the library. However, it is not our core business and turnaround time on issus or PR reviews might be longer than usual due to other priorities. We do not offer paid consulting services around this library.

Usage

Parsing:

from drafthorse.models.document import Document
samplexml = open("sample.xml", "rb").read()
doc = Document.parse(samplexml)
print(doc.trade.agreement.seller.name)

Document.parse() taskes a boolean parameter strict which defaults to True. This means that the parser will raise an error if it encounters any unknown element. If you set it to False, the parser will not raise an error and parse whatever it can.

Generating:

from datetime import date, datetime, timezone
from decimal import Decimal

from drafthorse.models.accounting import ApplicableTradeTax
from drafthorse.models.document import Document
from drafthorse.models.note import IncludedNote
from drafthorse.models.party import TaxRegistration
from drafthorse.models.tradelines import LineItem
from drafthorse.pdf import attach_xml

# Build data structure
doc = Document()
doc.context.guideline_parameter.id = "urn:cen.eu:en16931:2017#conformant#urn:factur-x.eu:1p0:extended"
doc.header.id = "RE1337"
doc.header.type_code = "380"
doc.header.name = "RECHNUNG"
doc.header.issue_date_time = date.today()
doc.header.languages.add("de")

note = IncludedNote()
note.content.add("Test Node 1")
doc.header.notes.add(note)

doc.trade.agreement.seller.name = "Lieferant GmbH"
doc.trade.settlement.payee.name = "Lieferant GmbH"

doc.trade.agreement.buyer.name = "Kunde GmbH"
doc.trade.settlement.invoicee.name = "Kunde GmbH"

doc.trade.settlement.currency_code = "EUR"
doc.trade.settlement.payment_means.type_code = "ZZZ"

doc.trade.agreement.seller.address.country_id = "DE"
doc.trade.agreement.seller.address.country_subdivision = "Bayern"
doc.trade.agreement.seller.tax_registrations.add(
    TaxRegistration(
        id=("VA", "DE000000000")
    )
)

doc.trade.agreement.seller_order.issue_date_time = datetime.now(timezone.utc)
doc.trade.agreement.buyer_order.issue_date_time = datetime.now(timezone.utc)
doc.trade.settlement.advance_payment.received_date = datetime.now(timezone.utc)
doc.trade.agreement.customer_order.issue_date_time = datetime.now(timezone.utc)

li = LineItem()
li.document.line_id = "1"
li.product.name = "Rainbow"
li.agreement.gross.amount = Decimal("999.00")
li.agreement.gross.basis_quantity = (Decimal("1.0000"), "H87")  # H87 == pieces
li.agreement.net.amount = Decimal("999.00")
li.agreement.net.basis_quantity = (Decimal("999.00"), "EUR")
li.delivery.billed_quantity = (Decimal("1.0000"), "H87")  # H87 == pieces
li.settlement.trade_tax.type_code = "VAT"
li.settlement.trade_tax.category_code = "E"
li.settlement.trade_tax.rate_applicable_percent = Decimal("0.00")
li.settlement.monetary_summation.total_amount = Decimal("999.00")
doc.trade.items.add(li)

trade_tax = ApplicableTradeTax()
trade_tax.calculated_amount = Decimal("0.00")
trade_tax.basis_amount = Decimal("999.00")
trade_tax.type_code = "VAT"
trade_tax.category_code = "AE"
trade_tax.exemption_reason_code = 'VATEX-EU-AE'
trade_tax.rate_applicable_percent = Decimal("0.00")
doc.trade.settlement.trade_tax.add(trade_tax)

doc.trade.settlement.monetary_summation.line_total = Decimal("999.00")
doc.trade.settlement.monetary_summation.charge_total = Decimal("0.00")
doc.trade.settlement.monetary_summation.allowance_total = Decimal("0.00")
doc.trade.settlement.monetary_summation.tax_basis_total = Decimal("999.00")
doc.trade.settlement.monetary_summation.tax_total = Decimal("0.00")
doc.trade.settlement.monetary_summation.grand_total = Decimal("999.00")
doc.trade.settlement.monetary_summation.due_amount = Decimal("999.00")

# Generate XML file
xml = doc.serialize(schema="FACTUR-X_EXTENDED")

# Attach XML to an existing PDF.
# Note that the existing PDF should be compliant to PDF/A-3!
# You can validate this here: https://www.pdf-online.com/osa/validate.aspx
with open("input.pdf", "rb") as original_file:
    new_pdf_bytes = attach_xml(original_file.read(), xml)

with open("output.pdf", "wb") as f:
    f.write(new_pdf_bytes)

Document.serialize() will validate the generated XML against the specified schema and raise an error if it is not valid. If you want to avoid validation, you can set the schema parameter to None.

Development

To run the included tests:

pip install -r requirements_dev.txt
py.test tests

To automatically sort your Imports as required by CI:

pip install isort
isort -rc .

To validate files using mustang:

git clone https://github.com/ZUGFeRD/mustangproject.git
cd mustangproject
git checkout core-2.9.0
./mvnw clean package
java -jar Mustang-CLI/target/Mustang-CLI-2.7.4-SNAPSHOT.jar --action validate --source invoice.pdf

Security

If you discover a security issue, please contact us at [e-mail] and see our [Responsible Disclosure Policy](https://docs.pretix.eu/trust/security/disclosure/) further information.

Credits and License

Maintainer: Raphael Michel <[e-mail]>

License of the Python code: Apache License 2.0

The PDF handling (drafthorse/pdf.py) was originally based on the code of factur-x, Copyright 2016-2018, Alexis de Lattre <[e-mail]>, released under a BSD license.

The packages includes schemas and samples of the ZUGFeRD specification (.xsd and .xml files) which are owned by the Forum für elektronische Rechnungen bei der AWV e.V („FeRD“) and are released under a proprietary license that allows bundling them together with other software for free.